Practice Education Quality Evaluation Based on Parallel Genetic Algorithm Elman Neural Network
For the evaluation problem of practice education, the existing evaluation indexes are imperfect.Moreover, although the neural network has been used to evaluate these indexes, but the accuracy of evaluation algorithms remains to be improved.In this study, we first design five primary evaluation indexes and each primary index has its second index.Then we propose a parallel Elman network and use genetic algorithm to optimize the initial weights and thresholds in each subnet before the training process.For this reason, the proposed algorithm for practice education quality evaluation can be called PGAE (Parallel Genetic Algorithm Elman neural network).Based on these indexes, the results suggest that our PGAE algorithm is effective for the evaluation problem of practice education and the accuracy of PGAE algorithm is much better than the fuzzy network and the original Elman network.In addition, the mean square error of PGAE algorithm is 6.2e-003 which is smaller than another two neural network for an order of magnitude.It also demonstrates that the PGAE network has high information utilization, reasonable and efficient network structure, and intelligent recognition mode.Most importantly, this provides a new approach on the practice education quality evaluation and the PGAE algorithm also can be applied in other evaluate fields.
Practice education Quality evaluation Index Parallel genetic algorithm Elman neural network
JIN CUIHONG
College of Mechanical Engineering, Chongqing University, Chongqing, China
国际会议
The 11th International Conference on Modern Industrial Training(第十一届现代工业培训国际学术会议)
北京
英文
611-615
2015-10-23（万方平台首次上网日期，不代表论文的发表时间）